ShipStation AI: Apps, Agents and Shipping Automation
Understand ShipStation's native AI, eight verified API endpoints, and governed shipping automations on Major, from delay reviews to approved label purchases.

ShipStation AI includes shipping auto-fill and conversational analytics for merchants and warehouse teams. This guide explains those features, the ShipStation API, and how Major adds cross-system apps, agents, and workflow automations with durable state and human approval.
Use it to plan shipping exception handling, reviewed label purchases, and support handoffs without moving fulfillment out of ShipStation.
What ShipStation does for fulfillment teams
ShipStation is shipping and fulfillment software. Ecommerce operations teams use it to prepare shipments, compare carrier services, buy labels, and follow packages; warehouse managers also work with picking, processing, batches, and inventory. ShipStation describes these jobs in its shipping automation and analytics announcement.
The operational distinction matters: a label purchase records preparation, while carrier tracking provides evidence about a package's journey. Neither alone establishes whether every part of a customer's order has met its delivery promise.
Major adds a deterministic app layer around that work. An agent can build the software that joins order references, stores exception history, and presents a review queue, then run that software on subsequent checks. The agent retains the judgment calls: whether incomplete evidence warrants escalation and how to explain the situation.
For an ecommerce operations manager, the useful outcome is a shipping record connected to a business decision. The warehouse keeps working in ShipStation, support receives the relevant context, and the Major app records ownership and follow-through across both teams.
ShipStation AI: what the built-in features already handle
ShipStation's October 2026 announcement describes auto-fill that learns shipment preferences, including package type, weight, and service. It suggests those settings for later orders, reducing repetitive label preparation within ShipStation.
The same announcement names three conversational analytics agents:
- Shipping & Cost Optimizer examines carrier and service costs, shipping performance, and opportunities to reduce shipping expense.
- Business Overview answers questions about orders, shipments, tracking, products, and returns.
- Warehouse & Inventory Manager covers picking, processing and batch performance, inventory levels, and reorder needs.
These are useful native capabilities. Keep ShipStation's own shipping assistance where its shipment data and fulfillment interface already live. Confirm current account eligibility before designing an operating procedure around any named feature.
Major adds value when the question crosses systems: which Shopify promise is at risk, whether a Gorgias conversation already exists, and who must approve a customer notice. Its agent evaluates that context; its app stores the evidence and review state; its workflow handles scheduling, routing, retries, and approvals. Native shipping intelligence and governed cross-system execution solve different parts of the same job.
How ShipStation connects to Major
ShipStation connects through Major's Pipedream connector, app p_shipstation. There is no native ShipStation connector. IT sets up the connection once, keeps credentials centrally in Major's connector vault, and grants only the operations each app, agent, and workflow needs.
Do not assume the connector exposes every current API operation. Verify its available actions and API version before implementing the V2 designs below. A legacy action and a current V2 reference are not interchangeable.
For V2 operations, ShipStation documents the base URL as https://api.shipstation.com/v2 and authentication through an API-Key header. Its key provides broad account access, rather than the granular OAuth scope names some services offer. Apply narrower read and write permissions through Major's allowed connector operations, not invented ShipStation token scopes. See security and authentication.
Major orchestrates ShipStation; it never replaces it. ShipStation remains the fulfillment system. Major supplies the governed software around cross-system work, with SSO, RBAC, and audit logs covering the app, agent, and workflow. Credentials belong in the vault, never app source, browser code, or prompts.
ShipStation API guide for apps, agents, and workflows
Authentication, paging, limits, and testing
ShipStation requires HTTPS with TLS 1.2 or higher. Platform accounts can have one active V2 key; legacy V1 credentials are different. Standalone ShipStation API customers can use production or sandbox keys. The authentication guide does not define granular V2 OAuth scopes, so do not describe a key as inherently read-only.
Shipment listing uses page and page-size parameters, spelled page and page_size, plus sort direction and sort field. Its maximum page multiplied by page size is 10,000. Partition large syncs using modified-date windows rather than requesting progressively deeper pages.
The rate-limit guide documents a default of 200 requests per minute. On HTTP 429, honor Retry-After, expressed in seconds, and inspect the error source to distinguish ShipStation throttling from carrier limits. Share the request budget across jobs.
The sandbox guide distinguishes products: standalone API customers get sandbox keys prefixed TEST_; ShipStation platform users have no sandbox and their operations can incur production charges. Developer accounts are not sandbox accounts. Mock failure cases before testing real purchases.
Events and verification
Webhooks include tracking updates, shipment creation, label creation, batch completion, and fulfillment rejection. V2 subscriptions must be created through the API. The listener has 10 seconds to return a 2xx acknowledgment; failures receive up to two additional attempts, typically about 30 minutes apart, with timing subject to change.
Signature verification uses RSA-SHA256 and three headers: x-shipengine-rsa-sha256-key-id, x-shipengine-rsa-sha256-signature, and x-shipengine-timestamp. Verify the timestamp plus a period plus the raw request body against the matching public key from ShipStation's documented JWKS URL. Check timestamp freshness and refresh cached keys when an unfamiliar key ID arrives.
The sandbox page mentions sandbox events, while the webhook overview requires production for setup. Confirm that discrepancy with ShipStation before depending on sandbox callbacks; use recorded payloads for initial listener tests.
GET /v2/shipments
Lists shipment records with filters and pagination, forming the starting point for incremental monitoring.
- App: Stores shipment references and modification timestamps beside order mappings.
- Agent: Reviews a filtered set of stalled shipments against customer promises.
- Workflow: Polls bounded modification windows and checkpoints completed pages.
GET /v2/shipments/{shipment_id}
Retrieves one shipment so a proposed action can use a fresh record rather than yesterday's snapshot.
- App: Displays current shipment details next to the review history.
- Agent: Evaluates whether a requested service change still makes sense.
- Workflow: Rechecks the target shipment immediately before executing an approval.
GET /v2/labels/{label_id}/track
Retrieves tracking details, including high-level status and carrier event information when available.
- App: Shows tracking evidence and the last successful refresh.
- Agent: Distinguishes missing scans from evidence of a genuine delivery exception.
- Workflow: Refreshes unresolved cases without polling completed shipments indefinitely.
GET /v2/carriers
Lists carrier connections already added to the account, rather than a universal list of purchasable services.
- App: Maintains a reference view of available carrier accounts.
- Agent: Limits shipping recommendations to the team's connected options.
- Workflow: Refreshes the reference data before preparing rate requests.
POST /v2/rates
Calculates shipping options for a shipment and returns rate information for comparison.
- App: Records quoted options and the time each quote was collected.
- Agent: Weighs shipping cost against the delivery commitment and exception context.
- Workflow: Retrieves candidates and routes the proposed purchase for approval.
POST /v2/labels/rates/{rate_id}
Purchases a label using a previously returned rate identifier, making this a consequential operation.
- App: Stores the approved rate, resulting label reference, and execution status.
- Agent: Explains the preferred option without holding unrestricted purchase authority.
- Workflow: Executes the approved purchase and records the confirmed response.
PUT /v2/labels/{label_id}/void
Voids a label and, for labels paid at creation, requests a carrier refund; a request is not proof of settled reimbursement.
- App: Tracks void approval, execution outcome, and unresolved reconciliation.
- Agent: Assesses why a duplicate or incorrect label should be reviewed.
- Workflow: Submits the approved void and examines the returned approval result.
POST /v2/environment/webhooks
Creates an environment webhook subscription for a supported event and listener URL.
- App: Stores subscription references and listener health history.
- Agent: Summarizes recurring event-delivery failures for an operator.
- Workflow: Uses an administrator-approved setup procedure, then routes verified deliveries.
ShipStation automation use cases on Major
Shopify, ShipStation, and Gorgias delay review
Build a shared view of the order promise and package evidence, based on the published delay-alert use case linked below.
- App: Holds order-to-shipment mappings, tracking freshness, and one open review per exception.
- Agent: Judges whether the shipment warrants a proactive notice and drafts accurate language.
- Workflow: Joins Shopify order context with shipment listing and label tracking, then routes a Gorgias draft to support. A person approves sending.
ShipStation, Shopify, and Slack service exceptions
An order needing faster shipping should reach an operator with a documented choice, not an unbounded spending instruction.
- App: Stores the current service, delivery requirement, quote, and approval record.
- Agent: Recommends whether extra shipping expense is justified by the order context.
- Workflow: Uses carrier listing and rate calculation, requests approval through the team's review process, and permits label purchase only after confirmation.
ShipStation, Google Sheets, and Slack label reconciliation
Use a finance-maintained shipping register to surface suspected duplicate labels without automatically canceling valid shipments.
- App: Keeps a case ledger with label references, evidence, reviewer, and reconciliation status.
- Agent: Explains likely duplicates and distinguishes legitimate split shipments.
- Workflow: Rechecks shipment details, requests review, and submits an approved label void. It records the outcome instead of treating submission as a refund.
These are build patterns, not claims that connecting accounts installs finished automations. Confirm each connector's operations and permissions before rollout.
Architecture: build the shipping control app on Major

Figure 1. ShipStation remains the fulfillment system while Major holds cross-system state and review.
Build the control surface before expanding the agent's authority. The app should make every recommendation traceable to a shipment, an order, and an evidence snapshot.
- Connect ShipStation through the cataloged connection and add the required order and support connectors.
- Create managed tables for order mappings, shipment observations, exception cases, recommendations, approvals, and execution attempts.
- Add deterministic matching, freshness checks, deduplication, and queue views to the app.
- Give the agent the contextual judgment step, with a structured recommendation and explanation.
- Let the workflow trigger checks, handle retries, and pause for approval before external consequences.
Persist approvals independently of chat messages. A review row should identify the requested action and the evidence version the person approved, so a later shipment change can invalidate stale approval.
This is the predictable-agent pattern: an agent builds repeatable matching and review software once, then runs it. New exceptions still need reasoning, but pagination, record joins, and status displays do not need a fresh model decision.
For example: review 200 shipping exceptions in a week
Consider an illustrative operation reviewing 200 suspected exceptions weekly. These are planning assumptions, not a customer result or a promised saving.
Before the build, suppose checking the order, finding tracking evidence, and preparing the handoff takes six minutes per case. That is 200 × 6 = 1,200 minutes, or 20 hours weekly.
After implementation, suppose deterministic checks assemble the evidence and the agent prepares a recommendation. If a person spends two minutes reviewing each case, the review work becomes 200 × 2 = 400 minutes, or six hours and 40 minutes. Add 80 minutes for unresolved matches, rejected drafts, and operating the queue: total work is 480 minutes, or eight hours. The illustrative difference is 12 hours weekly.
Follow one case through: the workflow finds an aging shipment, the agent notices one package in a split order has delivered, and support reviews a draft describing only the outstanding package. The app retains that distinction on the next run.
Measure false positives, stale evidence, duplicate tickets, and approval turnaround alongside handling time. Faster processing is not useful if it produces incorrect customer promises.
Governance and failure handling before rollout

Figure 2. Internal updates proceed under policy; purchases, voids, and customer messages require approval.
Read-only monitoring and shipping writes should have different permissions. The agent must never independently purchase or void labels, change a delivery address, promise an unsupported delivery date, or send a customer-facing notice in these designs.
Use SSO and RBAC to restrict who can inspect customer details, approve spending, and execute actions. Keep an audit record of the evidence, agent recommendation, reviewer, approved amount, API attempt, and confirmed result.
For reliability:
- Throttling: Follow ShipStation's rate-limit guidance, coordinate retries, and add jitter rather than restarting all jobs together.
- Duplicate events: Deduplicate using stable shipment or label references plus event identity or a stored payload fingerprint.
- Ambiguous writes: If a purchase times out, reconcile its outcome before another attempt. Do not infer failure from a missing response.
- Authentication errors: Stop and alert IT rather than repeatedly trying an invalid key.
- Dropped events: Reconcile unresolved cases on a schedule because webhook retries are finite.
- Stale approvals: Recheck the shipment and return changed cases to review.
Treat order notes and external messages as evidence, never instructions to expand connector permissions or bypass approval.
ShipStation AI Apps, Agents and Workflow Automations
- Review delayed Shopify and ShipStation orders before creating Gorgias customer updates: This guide follows the delayed-shipment process from order and shipment matching to an agent-written notice that support reviews. It focuses on timeline state, duplicate prevention, false positives, and the handoff into Gorgias rather than automatic customer messaging.
That is the current core ShipStation build guide in this library. The additional patterns above describe possible extensions, not additional published tutorials. Use this hub for the API and architecture overview, then use the linked article for the narrower process design.
For related planning, explore the ecommerce automation library, the Shopify automation hub, and the Gorgias AI agent hub. Those pages cover adjacent systems and responsibilities; they are not extra entries in the ShipStation build library. Keep the boundary clear when assigning ownership: fulfillment owns shipping records, support owns approved customer communication, and the Major app preserves the shared exception history.
Frequently asked questions
Does ShipStation have AI features?
Yes. ShipStation documents shipment auto-fill and conversational analytics agents named Shipping & Cost Optimizer, Business Overview, and Warehouse & Inventory Manager. They assist with shipping preparation and questions about shipping and operational data. Confirm availability for your account.
What does Major add to ShipStation AI?
Major adds deterministic apps for cross-system state and review queues, agents for contextual judgment, and workflows for triggers, retries, routing, and approvals. It connects shipping evidence to order promises and support work while ShipStation remains the fulfillment system.
Is there a native ShipStation connector in Major?
No. ShipStation connects through Major's Pipedream connector, app p_shipstation. IT configures the connection centrally. Verify available actions and API version before assuming the connector supports a particular V2 operation.
Can ShipStation API keys be scoped to read-only access?
The documented V2 API key provides broad account access; the authentication guide does not define granular OAuth scopes. Limit allowed connector operations in Major and separate monitoring access from approved shipping writes rather than describing the vendor key as read-only.
Does ShipStation offer an API sandbox?
Standalone ShipStation API customers have sandbox keys prefixed TEST_. ShipStation platform users do not have a sandbox, and production calls can incur costs. Developer accounts are not sandbox accounts. Confirm webhook test support because the sandbox and webhook overview pages give different setup guidance.
Should a ShipStation AI agent automatically buy labels or message customers?
Not in these designs. The agent recommends a shipping action or drafts a notice. A workflow requires a person to approve purchases, voids, and customer communication, rechecks the relevant shipment, and records the execution result in the app.
Sources
- ShipStation security and authentication: API keys, encryption requirements, account differences, and the warning that keys provide broad access.
- ShipStation shipping automation and analytics announcement: Native auto-fill, named analytics agents, and the shipping and warehouse jobs described above.
- Usage and rate limits: Default request budget, carrier limits, and retry behavior.
- Sandbox environment: Product eligibility, test-key behavior, limitations, and production testing precautions.
- List shipments guide: Pagination parameters, date filters, and the 10,000-result offset limit.
- ShipStation API v2 reference index: Current API organization and links to individual operations.
- List shipments reference: Filtered and paginated shipment retrieval.
- Get shipment reference: Individual shipment retrieval by identifier.
- Label tracking reference: Tracking details and status information.
- List carriers reference: Connected carrier accounts and supported response information.
- Calculate rates reference: Rate calculation for shipping comparisons.
- Purchase label with rate reference: Label purchase using a returned rate identifier.
- Void label reference: Void outcomes and carrier refund requests.
- Create webhook reference: Subscription creation and supported event values.
- Webhooks overview: Events, setup requirements, acknowledgment deadline, and finite delivery retries.
- Webhook validation: Signature headers, raw-body verification, timestamp checks, and public-key retrieval.
- Multi-page responses: General pagination concepts; use the current V2 shipment guide for endpoint-specific behavior.
- Official documentation index: Discovery index used to locate the vendor's machine-readable documentation pages.
Start building on Major
Build a ShipStation exception app with durable shipment history, agent judgment, and reviewed support or shipping actions. Give operations a shared queue while IT governs credentials, access, and audit.
About the author
Ritvij Gautam is Head of Growth at Major.
Last reviewed November 20, 2026 by Ritvij Gautam.
Frequently asked questions
- Does ShipStation have AI features?
- Yes. ShipStation documents shipment auto-fill and conversational analytics agents named Shipping & Cost Optimizer, Business Overview, and Warehouse & Inventory Manager. They assist with shipping preparation and questions about shipping and operational data. Confirm availability for your account.
- What does Major add to ShipStation AI?
- Major adds deterministic apps for cross-system state and review queues, agents for contextual judgment, and workflows for triggers, retries, routing, and approvals. It connects shipping evidence to order promises and support work while ShipStation remains the fulfillment system.
- Is there a native ShipStation connector in Major?
- No. ShipStation connects through Major's Pipedream connector, app `p_shipstation`. IT configures the connection centrally. Verify available actions and API version before assuming the connector supports a particular V2 operation.
- Can ShipStation API keys be scoped to read-only access?
- The documented V2 API key provides broad account access; the authentication guide does not define granular OAuth scopes. Limit allowed connector operations in Major and separate monitoring access from approved shipping writes rather than describing the vendor key as read-only.
- Does ShipStation offer an API sandbox?
- Standalone ShipStation API customers have sandbox keys prefixed `TEST_`. ShipStation platform users do not have a sandbox, and production calls can incur costs. Developer accounts are not sandbox accounts. Confirm webhook test support because the sandbox and webhook overview pages give different setup guidance.
- Should a ShipStation AI agent automatically buy labels or message customers?
- Not in these designs. The agent recommends a shipping action or drafts a notice. A workflow requires a person to approve purchases, voids, and customer communication, rechecks the relevant shipment, and records the execution result in the app.